Central Nervous and Other Systems in Vitro and in ...

Lactotetraose Series Ganglioside 3′,6′-isoLD1 in Tumors of Central Nervous and Other Systems in Vitro and in Vivo C. J. Wikstrand, D. C. Longee, R. E. McLendon, et al. Cancer Res 1993;53:120-126.

Updated version

E-mail alerts Reprints and Subscriptions Permissions

Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/53/1/120

Sign up to receive free email-alerts related to this article or journal. To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at [email protected]. To request permission to re-use all or part of this article, contact the AACR Publications Department at [email protected].

Downloaded from cancerres.aacrjournals.org on June 2, 2013. © 1993 American Association for Cancer Research.

[CANCER RESEARCH 53. 120-126. January I. 1993]

Lactotetraose

Series Ganglioside 3',6'-isoLD1 in TVimors of Central Nervous and

Other Systems in Vitro and in Vivo1 C. J. Wikstrand, D. C. LongÃ©e,2R. E. McLendon, G. N. Fuller, H. S. Friedman, P. Fredman, L. Svennerholm, D. D. Bigner3

and

Departments of Pathology 1C. J. W., D. C. L. R. E. M.. G. N. F.. H. S. F.. D. D. B.\ and Pediatrics Â¡H.S. F.Â¡.and Preuss Laboratory for Brain Tumor Research Â¡H.S. F., D. D. B. l, Duke University Medical Center, Durham, North Carolina 27710; and Department of Psvchiatry and Neitrochemislry, University of GÃ¶teborg. GÃ¶teborg,Sweden Â¡P.F.. L S.I

ABSTRACT

normal tissues remains; this is especially crucial within the CNS, a "ganglioside-rich" environment (7). For this reason, the discrimina

Two monoclonal antibodies, DMAb-21 and DMAb-22, directed against the lactotetraose series ganglioside-associated epitope IV-'NeuAc,!!!'1NeuAcLcOse4Cer i.V.h'-isol.,,, i. were found to define the minimum bind

tion of CNS tumor-associated gangliosides with negligible expression by components of the normal adult neuraxis as eventual targets for compartmental, intrathecal administration has been actively pursued. We and others have described the expression of the lactotetraose series gangliosides IV'NeuAcLcOsCjCer (3'-isoLMI) and IV3NeuAc,IH6NeuAc-LcOse4Cer (3',6'-isoLD,) by human tumors, derived xeno grafts. and fetal tissues (7, 8-12); the presence of the 3',6'-isoLD,

ing epitope NeuAc(or NAc. The distribution

NeuGc)a2-3GalÃŸl-3(NeuAc or NeuGc)a2-6Glcof 3',6'-isoLDi in cultured cell lines and derived

xenografts of primary tumors of the human central nervous system and of embryonal or neuroectodermal tumor derivation was determined. Only 4 of 26 cell lines, 3 teratomas and 1 pancreatic adenocarcinoma, expressed detectable .V.d'-isol.,,, when cultured in vitro: none of 14 tested glioma

epitope on undefined carrier molecules in the serum of human cancer patients has been reported (13). We have recently described a MAb specific for 3'-isoLM1 (SL-50) and investigated the distribution of this

lines, including 2 that expressed the monosialo-precursor IWuAcLcOse.tCer in vitro, expressed detectable levels. Expression of 3',6'isoL|>i was more frequent when neoplastic cells were grown in xenograft form in athymic mice; 4 of 10 glioma and 2 of 2 teratoma xenograft ganglioside extracts were positive for .V.(i'-iv>l,,,,. The absence of 3',6'-

antigen in human tumors of the CNS and in xenografts and cultured cell lines derived therefrom (12). The expression of both 3'-isoLM, and 3',6'-isoLDi within the human CNS is associated with periods of

iscil,,,, in cultured tumor cells of the central nervous system and its pro portional increased presence in tumor cells of the same origin grown in vivo further supports previous studies suggesting that ganglioside ex pression may be modified by environmental forces. The expression of laclo series gangliosides both in vitro and in viro by teraloma and pancreatic adenocarcinoma cells, as opposed to only in vivo expression by glioma cells, suggests that tissue-specific forces may also exist. Immunohistochemical localization of 3',6'-isoLm in frozen sections of primary central ner

intense astroglial proliferation occurring in the fetal human forebrain through the first trimester, gradually disappearing after the age of 2 years (14). The ready detection of elevated amounts of these gangli osides in the brain tissue of children succumbing to polyunsaturated fatty acid lipidosis as compared with undetectable levels in agematched (2-20 years) negative controls ( 15) has led to the suggestion that 3'-isoLM1 and 3',6'-isoLD, may be characteristic gangliosides of

vous system neoplasms including those of glial and nonglial origin was performed; 20 of 30 (67%) of glial tumors were positive. Among nonglial tumors, 21 of 34 (62%) of epithelial cancers were reactive with anti-3',6'-

proliferative astroglial cells. The distribution of 3'-isoLM,

in tumors of the CNS has been

approached through the use of the specific MAb SL-50 (12); the reactivity of DMAb-14 for both the monosialyl and disialyl forms precluded its definitive use in immunohistochemistry, thus the distri bution of 3',6'-isoLm in human tumors has not been directly ad dressed. We report here the isolation of two specific anti-3',6'-isoLD,

Â¡sol.,,, monoclonal antibodies; notably negative were carcinomas of the ovary and lung carcinomas of all subtypes. Lymphomas and infiltrative lymphocytes were uniformly negative. The restriction of 3'.6'-isol.,,, ex pression within the human central nervous system to periods of fetalneonatal astroglial proliferation, to intense reactive astrocytosis, and to primary neoplasms, and the production of specific monoclonal antibodies to this epitope provide a specific complex for immunolocalization and, eventually, immunotherapy.

MAbs and their application in both immunohistochemical and HPTLC analysis of human tumor tissues, derived xenografts, and cell lines. Similar to the distribution of its precursor, 3'-isoLM,, 3',6'-isoLD1 is not expressed in CNS tumor cell lines, but is detectable in over 50% of tumor biopsies and xenografts. The expression of 3',6'-isoLm by

INTRODUCTION Targeting of quantitatively increased "normal" ganglioside epitopes

teratomas and carcinomas further extends the prospective utility of this ganglioside as a target for immunolocalization.

on tumor cells has been shown to be effective for both localizing |G[)2 (1), GMJ (2)]4 and therapeutic [GD3 (3, 4); GD2 (5, 6)] applications.

MATERIALS

Despite these promising studies, the capacity for uptake by nontarget

AND METHODS

Glycolipids. Gangliosides und neutral glycolipids used as standards and references were isolated and characterized by fast atom bombardment-mass spectrometry. NeuAc-GM2 was isolated from Tay-Sachs brain (16); GM1 and

Received 8/12/92; accepied 10/20/92. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 Supported by NIH Grants CA 11898. NS 20023. T32-NS 07304, CA 56115. and CA 44640; by the Swedish Medical Research Council (Project 03X-09909-01); and by the National Swedish Board for Technical Development (Project 84-4667P). : Present address: Division of Hematology/Oncology, Arkansas Children's Hospital.

GJX were purified from metastatic melanoma tissue removed at surgery; and Gr>2was prepared by bovine /3-galactosidase treatment of Gun, purified from normal adult human brain. Purified enzyme was kindly provided by Dr. George Jourdian, University of Michigan, Ann Arbor. MI. 3'-isoLMi and GalNAc-3'iM'l M were purified from human meconium (17). 3'-LM[ and 3',8'-Lm from cauda equinadS), and 3'-isoLM, and 3',6'-isoLD, from human glioma cell line

800 Marshall Street. Little Rock. AR 72202-359I. 'To whom requests for reprints should be addressed, at Box 3156. Department of

D-54 MG-induced nude mouse and nude rat xenografts (8), or from brains of

Radiology. Duke University Medical Center. Durham. NC 27710. 4 The abbreviations used are; Ganglioside and gtycolipid designations are according to CBN recommendations (IUPAC-IUB Commission) and Svennerholm (39. 40): GD2. II1 (NeuAchGgOse,Cer; GM.i. II'NeuAc-LacCer; GD,, II'INeuAcbLacCer; 3'-isoLM,. IV'NeuAcLcOse4Cer; 3',6'-isoLm. IVNeuAcJI^NeuAcLcOse^er; GM2. Il'NeuAcGgOse,Cer; GD,h. II'INeuAcbGgOseaCer; 3'-LMi. IV'NeuAcnLcOse4Cer; 3'.8'-LDi. IV'(NeuAc);nLcOse4Cer; GMi. II'NeuAcGgOse^er; GDI., IV'NeuAc.II'NeuAcGgOse4Cer: GTii>. IV'NeuAcII1(NeuAc)2GgOse4Cer; iso-LAi. LcOse4Cer; CNS. central nervous system; MAb. monoclonal antibody; HPTLC, high-performance thin-layer chromatography; SP-RIA, solid phase radioimmunoassay.

patients dying of polyunsaturated fatty acid lipidosis (15). All other glycolipids were purified from normal adult human brain. The isolation and characteriza tion of gangliotetraose series gangliosides has been described (17, 19, 20). Monosialo- and oligosialoganglioside fractions were prepared from cultured cell pellets, normal tissue, or tumor xenografts of known weight and/or cell count as previously described (12). Densitometric scanning of resorcinolvisualized ganglioside bands was performed at 620 nm with a CAMAG TLC 120


GANGLIOSIDE

.V.6'-ÃŒM>LDi IN HUMAN

Scanner II (CAMAG. Multen/., Switzerland). Quantitative measurement of total ganglioside sialic acid was performed by the resorcinol assay (21). Cell Lines. The established permanent human glioma- and medulloblastoma-derived cell lines used in this study have been described (12, 22-27). Neuroblastoma cell lines SK-N-MC. SK-N-SH. LAN-1. and LAN-5 were kindly provided by Dr. Robert Seeger. UCLA. The American Type Culture Collection. Rockville, MD. was the source for cell lines TERA-1, TERA-2. SK-MEL-28. IMR-32. and P3X63/Ag8.653 (653). The cell line N4I7D (small cell lung carcinoma) was obtained from the National Cancer Institute; and the embryonal carcinoma cell line PA-1 was the gift of Dr. James Trosko, Mich igan State University, East Lansing. MI. The propagation, storage, and testing of these cell lines to ensure the absence of HeLa cell contamination, inter- or intra-cell line contamination, or Mycoptasma infection have been published elsewhere (28). Hybridoma Production. Female BALB/c mice, 15 weeks old (Charles River Breeding Laboratories. Stoneridge. NY), were used for immunization with PA-1 leratoma cells grown in xenografi form in athymic mice: PA-1cultured and xenograft cells express both 3'-isoLM, and 3'.6'-isoLm (10. 12). Donors received multiple doses of trypsini/.ed (0.25% ) PA-1 murine xenograft cells i.p.: fusion with 653 cells was performed by our standard procedure (29. 30). Initial reactivity screen was performed against mono- and oligosialoganglioside fractions of D-54 MG xenograft cells, of which 3'-isoLM, and 3',6'isoL[,i are the predominant mono- and disialogangliosides. respectively, to optimi/e selection for the target ganglioside 3',6'-isoLD1, with minimal selec tion of hybrids producing antibodies to other PA-1-associated gangliosides. Two hybrids (DMAb-21 and DMAb-22) were selected tor further study on the basis of liter, stability, and apparent specific ganglioside binding, determined by comparative immunostain of 3'.6'-isoLD1-containing extracts as defined by DMAb-14 (12). Hybrids were cloned in methylcellulose semisolid medium (29) three times and cultured for the production of antibody-containing super

slides, acetone fixed at -20Â°C. and air dried. Rehydrated sections were blocked with 10% normal goat serum, sequentially exposed to appropriate primary reagents, and developed with the Zymed biotinylated antibody-streptavidindiaminobenzidine system (33). Slides were counterstained with hematoxylin. mounted, and scored for reactivity by a minimum of two observers. The tissues were evaluated microscopically for immunoreactivity in both normal and neoplastic cells. Positivity was evaluated with respect to nuclear, cytoplasmic. and/or membranous localization of bound reagents. Each neoplasm was graded with regard to percentage of neoplastic cells positive on a tour-tiered scheme: less than 25% cells positive (Â±), 25-50% cells positive (+), 50-75% cells positive (++). and 75-100% (+++) cells positive. Any antigen localization in benign cells within the tissue section was noted.

RESULTS Reactivity and Specificity of DMAb-21 and DMAb-22. The two hybridomas. DMAb-21 and DMAb-22, were initially selected for further analysis on the basis of positive activity for D-54 MG rat xenograft oligosialoganglioside fractions by SP-RIA and HPTLC im munostain in comparison with DMAb-14 (Figs. 1 and 2). As revealed by the titration in SP-RIA of purified MAbs against the total oligo sialoganglioside fraction of D-54 MG rat xenograft. no difference in the activity of the three MAbs is apparent. However, separation of the components of the oligosialoganglioside extract by HPTLC in chloroform:meihanol:0.253 is considered positive as these values exceed the mean background value by >3 SDs (32). For SP-RIA. known PM amounts of ganglioside in methanol were

rj~o-a-i

GANGLIO SERIES GM2

plated in Polyvinylchloride microtiter plates (Dynatech Laboratories, Inc.. Chantilly. VA) and allowed to dry by evaporation. Before immunoassay. block ing of nonspecific binding was performed by a 30-min exposure at room temperature to incubation buffer (50 UMTris-HCI, 15 HIMNaCl, 3% low-fat milk, pH 7.4). Incubation times at 37Â°Cwere performed for 2 h; results were

r

GD2

calculated as for cell surface radioimmunoassay. HPTI.C Immunostain. Gangliosides. either purified to homogeneity or as mono- or oligosialoganglioside fractions extracted from cultured cell pellets,

ami

O-O-O-B..

Â«Mb

Dâ€”o~a~u-

GT1b lACTOSYL SERIES

normal human brain, or xenograft tissues, were separated by HPTLC, and reactivity with individual MAbs was determined as previously described (12. 34). Reference gangliosides were visualized with orcinol reagent (0.1% orcinol in distilled H2O. 3% H2SO4). Immunohistochemical Staining of Frozen Tumor Tissue. Immunohistological procedures have been thoroughly described (12. 33): snap-frozen spec imens, maintained at -135Â°C, were cut 5- to l()-um thick, applied to glass

GD3 GALACTOSE N-ACETYLGLUCOSAM1NE

|

r

ti AU I Yl MuÂ« AMIMI

Ai in

^

N GLYCOLYLNEURAMINIC

O

N ACETYIGAIACTOSAMINE

ACID

FUCOSE _

pt-3 LINKAGE

~

Â» 4 UN1(*GE

Fig. 1. Specificity analysis with defined gangliosides/glycolipids.

121


GANGLIOSIDE

.V.6'-isoLm IN HUMAN

ANTI-3',6'lSO-LDlMabs

ganglio series (GM2, GD2, GD,a, Gmh), the lactosyl series (GD,), and the neolactotetraose series (3'-LM,, 3',8'-LDI) are not bound by these MAbs. Unlike DMAb-14. both DMAb-21 and DMAb-22 require sia lylation of the subterminal /V-acetylglucosamine. Fucosylation of this sugar, as in Fuc-3'-isoLM, (Fig. 1), does not permit binding of

â€”¿ GM2 â€”¿ GM1

DMAb-21 or DMAb-22. Therefore, the minimum binding epitope of these MAbs is suggested to be NeuAc(or NeuGc)a2-3Gal/31-3(NeuAc or NeuGc)a2-6GlcNAc. Distribution of 3',6'-isoLâ€ž, in Cell Lines and Xenografts of

â€”¿ GDb â€”¿ GDlb â€”¿ GTlb 123

123

DMab-14

l:3'ISO-LMl

I

2

Human Tumors. Oligosialoganglioside fractions were prepared from packed cell pellets or xenograft tissue as described in "Materials and Methods." Human cell lines examined included 14 gliomas, 5

3

DMab-21 DMab-22

2:3',6'ISO-LD1

3

TUMORS

medulloblastomas, 3 teratomas, and 5 lines of various origins: 1 pancreatic carcinoma (HPAF), 1 small-cell carcinoma of the lung (N417). 1 melanoma (SK-MEL-28), 1 rhabdomyosarcoma (TE-671), and 1 neuroblastoma (LAN-1) (Table 1). Of these cell lines, 10 glio

Reference Standard

mas, 5 medulloblastomas. 2 teratomas, and 3 of the miscellaneous lines had corresponding murine xenografts from which oligosialogan glioside fractions were obtained. HPTLC immunostain analysis of these ganglioside extracts was performed with DMAb-21 and DMAb22; complete concordance of reactivity was observed. The results of these assays are summarized in Table 1 and illustrated in Fig. 4. Four of 26 cell line extracts examined expressed 3',6'-isoLD|: all 3 ter-

Fig. 2. HPTLC analysis of anti-3'.6'-isoLD, MAbs DMAb-14. DMAb-21. and DMAb-22 versus D-54 MG ral xenngraft monosialo- (Lane I} and oligosialoganglioside (Lane 21 fractions. Lane j consists of Ihe purified ganglioside standards listed on ordinate; total ganglioside sialic acid concentrations are. Lane I. 500 pmol: iMne 2, KKX)pmol; Lane ), 2000 pmol.

lack of recognition of GM2, GM,, GDia, GD|h, or GT,h. The specificity of the reactivity of these MAbs is further demonstrated in Fig. 3, illustrated by DMAb-22. Purified 3',6'-isoLD, isolated from D-54

atoma cell lines and the pancreatic carcinoma cell line HPAF. No glioma or medulloblastoma cell line extract expressed 3',6'-isoLm; the glioma cell line extracts of D-259 MG and D-263 MG, which express 3'-isoLM, (12), did not express the disialo derivative. Six of

MG rat xenografts is readily detected in as little as 1 pmol concen trations; the reactivity of DMAb-22 for 3',6'-isoLD, in murine D-54

19 xenograft ganglioside extracts examined expressed detectable 3',6'-isoLm: 4 of 10 glioma xenografts and 2 of 2 teratoma xeno grafts. All positive xenografts also expressed 3'-isoLM, (Ref. 12: data

MG xenografts (Lane A ) as compared with the rat xenograft fractions (Lanes B and C) reflects both differences in ceramide composition and preponderance of /V-glycolyl versus N-acetyl sialylation in mice as opposed to rats (12). No other ganglioside was detected in these fractions. Oligosialoganglioside extracts from human cauda equina (Lane D), human 20-week-gestation fetal brain (Lane E), and brain tissue from a case of polyunsaturated fatty acid lipidosis (Lane F ) also contained 3',6'-isoLDI, but in much lower concentrations than in D-54

not shown). Immunohistological Evaluation of Human Neoplasms. Both DMAb-21 and DMAb-22 were screened against the panel of frozen, acetone-fixed tissues listed in Table 2: again, results with the two antibodies were completely concordant; 15 of 30 of the glial tumor blocks were tested repetitively on consecutive sections with no significant discrepancies noted in the antigen localization patterns. Immunohistochemical staining with both DMAb-21 and DMAb-22 revealed 14 of 23 gliomas to exhibit localization in the neoplastic cell population (Â±to +++). Reactivity with the MAbs roughly correlated with glioma cell size; cells with abundant cytoplasm tended to exhibit strong cytoplasmic immunoreactivity and cells with scant cytoplasm tended to be nonreactive (Fig. 5a). A giant cell glioblastoma and a glioblastoma multiforme with large populations of gemistocytes

MG xenografts (2X, 5X, and 10X concentrations, respectively, of extract required for detection). DMAb-22 does not react with purified 3',8'-L,)i from human RBC (Lane G). the oligosialoganglioside frac tion of normal adult brain cerebral cortex (Lane H), or purified Fuc-3'-isoLM, (the CA-50 antigen, Lane I). The results of these analyses with purified, fast atom bombardmentmass spectrometry verified ganglioside and glycolipid standards are summarized in Fig. 1. The reactivity spectra of DMAb-21 and DMAb-22 are similar to that of DMAb-14 in that gangliosides of the

~

B

D

H

pmol 3'6-isoLDI Fig. 3. Binding of DMAh-22 to purified 3',6'-isoLDi and to this antigen in oligosialoganguoside fractions from various tissues. Extract preparation and HPTLC assay are described in "Materials and Methods"; DMAb-22 concentration is 5 ug/ml. The plate was chromatographed in chloroform:methanol:0.25'Â£ KCl (50:40:10, v/v/v). The amount of ganglioside sialic acid of the oligosialoganglioside extracts spotted/lane are given. Lane A, D-54 MG. murine xenograft. 500 pmol; Lanes B and C two preparations of D-54 MG rat xenografts. 500 pmol; Ulne D. human cauda equina. 1000 pmol; Lane E. 20-week-gestation human fetal brain. 2500 pmol; Lane F, polyunsaturated fatty acid lipidosis brain tissue. 5000 pmol; Lane C, 3'.8'-Lr>i purified from human RBC. 500 pmol; Lane H, human adult brain tissue. 3200 pmol; Lane I, Fuc-3'-isoLMi (CA-50 antigen; IV'NeuAc.III4Fuc-LcOse.i-Cer): Lanes I, 2. and 4. 3',6'-isoLu, purified from D-54 MG rat xenografts. 1, 2, and 4 pmol. 122


CANCil.lOSIDI-

Table I HPTLC iminunostain analysis of oligosialoganglioside crÃ¬ilines and \enografts

r.ft'-ivil.ni

extracts of cultured

Malignant gliomas D-32cl2 MG, D-37 MG. U-105 MG. U-410 MG, D-247 MO. D-270 MG

cytoplasm of many of the tumor types examined. Membrane reactivity was never identified without concomitant cytoplasmic reactivity. The majority (629^) of epithelial cancers studied were reactive with anti3',6'-isoLm MAbs; notable exceptions being carcinomas of the ovary

Xenograft

Cell culture

NA" 0/7

U-118 MG, U-251 MG. D-344 MG. D-245 MG D-212 MG. D-298 MG

NA 0/4

D-54 MG, D-259 MG. D-263 MG D-256 MG

(0 of 3 positive) and lung of all types (2 of 7 positive). A common feature in carcinomas was the presence of antigen localization also in the benign desmoplastic fibrous connective tissue about the infiltrat ing neoplasm, a pattern independent of neoplastic cell localization. As demonstrated in Fig. 6. teratomas exhibited MAb localization in the immature mesenchymal elements, a small round cell element, the immature basaloid epithelial cells of a hair shaft component, and faint positivity in a mature neural element (Fig. 6. a and /;). Intense staining of neoplastic cells is evident in 3 of 3 leiomyosarcomas studied (Fig. 6, c and it), while the breast carcinomas revealed punctate intranuclear, possibly nucleolar. antigen localization, a pattern not seen in any other neoplasm (Fig. 6f ). No antigen localization occurred in lymphomas: the absence of lymphoid cell staining is apparent in a germinoma tissue section in which the large neoplastic cellular com ponent is prominently stained, and the infiltrative lymphocyte com ponent is totally negative (Fig. 6/).

0/6

NA

on

Total malignant gliomas Medulloblaslomas Daoy. D-283 Med. D-341 Med. D-384 Med. D-425 Med Total medulloblastomas Teralomas and carcinomas PA-I.TERA-2 TERA-I. HPAF N-417 Total leratomas and carcinomas

0/14

4/4 4/10

0/5

0/5

+ NA 4/5

Miscellany TE-671 (rhabdomyosarcoma) SK-MEL-28 (melanomal LAN-1 (neuroblastoma) Total miscellany

IVMORS

ger than the cytoplasmic pattern. Nuclear reactivity was evident in three glioblastomas multiforme. Results of analysis of nonneural neoplasms are also presented in Table 2. Both DMAb-21 and DMAb-22 bound the membranes and

Positive immunostain with DMAb-21 and DMAh-22" Cell line designation

IN HUMAN

2/3

NA

oa

Table 2 SumirÃ-an-of DMAb-21 and DMAb-22 reactivity with acetone-fixed, frozen

0/2

sections of human tumor tissues 4/26 6/19 Total all categories " Reactivity to both DMAb-21 and DMAb-22: there was no detectable discrepancy

tumorsTypeAnaplaslic Glial casespositive of "/no.of

between the reactivity of the two M Abs. Up to 2 nmol ganglioside sialic acid were loaded per lane, depending upon the optimal concentration tor band development; all extracts testing negative were tested at 4 nmol/lane. representing as much as 48(M) (D-54 MG cells), or 36(K) (D-259 MG cells) ug sample protein/lane. For samples demonstrating positivity. the range of sample protein loaded was 950-26IX) ug protein/lane. * NA. not available.

casespositive/no.of of

casesexamined4/413/23212I/INonglialTypeEpithelial casesexamined21/345/63/33/54/60 astrocytomaGlioblastoma multiformeGliosarcomaGiant

cancersBreastProstateColonRenal

cell glioblastomaNo. cellOvaryPancreasLung

exhibited strong reactivity (+++) (Fig. 5, b and d). Fibrillar glioblastomas multiforme exhibited moderate, diffuse cytoplasmic reactivity (+++) of neoplastic cells. Glioblastomas multiforme composed of predominantly small cells, also known as "anaplastic cells" (35) were

cell)Lung (small ''TeratomasGerminomasSeminomasMelanomasLeiomyosarcom (other)

negative. In heterogeneous tumors, or tumors with mixed small and non-small cell populations, the non-small cell component was prom inently stained as compared to the negative small cell component (Fig. 5r). Cytoplasmic reactivity was the predominant reactivity pat tern in the glioblastoma multiforme; membranous localization was never observed without cytoplasmic localization, and was never stron

GLIOMA CELL LINESAND XENOGRAFTS

" Number of cases in which the neoplaslic cell component was determined positive. ''Three cases consisting of one moderately differentiated adenocarcinoma. one large cell carcinoma, and a squamous cell carcinoma that was weakly positive.

PAN ADCA TERATOMAS J i

MISCELLANY

iHn

-GM1 -GDla -GDlb -GTlb C X C L

l

I

X

l

l

2

X "

C X ""if"

X 5

X C X C J V 9

X 6

X 6

_ IO

Fig. 4. Oligosialoganglioside fractions of cultured cell pellets (C) and athymic murine xenografts (X) (unless otherwise designated as rat xenografls). Assays were performed as described in "Materials and Methods" by HPTLC immunostain. Unless otherwise noted. 2000 pmol of ganglioside sialic acid were loaded per lane. DMAb-22 concentration is 5 ug/ml. The plate was chromatographed in chloroform:methanol:0.25Ã-/c KCI (50:40:10. v/v/v) as above, migration of simultaneously assayed orcinol-developed reference standards appear on ordinale. Lanes I. U-251 MG; Lanes 2. D-54 MG; Lanes 3. D-259 MG; Lanes 4, D-263 MG; Lane 5, D-256 MG X only; Lanes 6. D-54 MG rat X; Lanes 7, PA-1; Lanes 8, TERA-2; Lane 9. HPAF; Lane NH COR. normal human cerebral cortex. 41XX)pmol; Lanes 10. SK-MEL-28. 123


GANOLIOSIDE

.V.6'-isoLD, IN HUMAN

TUMORS

Fig. 5. Reaclivily of DMAh-21 and DMAb-22 tor primary CNS tumors: immunohistochemistry of acetone-fixed, frozen sections. All MAbs were tested at 5 ug/ml. (a) TB 116. glioblasioma multiforme. DMAb-22. X 520. Diffuse pattern: intense staining of nearly all neoplastic cells, (b) TB 121. glioblastoma multiforme. DMAb-22. x 400. In tense staining of neoplastic gemistooytic cell bod ies and processes.( TB 65. gliosarcoma. DMAb22. X 170. Staining of neoplastic glial element and lack of staining of sarcomatouscomponent. (d ) TB 69. giant cell glioblastoma. DMAb-21. X 400. In tense staining of neoplastic giant cells.

S^S^m ^ ;. \'.* â€¢¿ --;i' d.'â€¢â€¢"â€¢ â€¢¿Â» â€¢¿. i